Research map / Graph learning / GNN architectures
Ewald-based Long-Range Message Passing for Molecular Graphs
Arthur Kosmala, Johannes Gasteiger, Nicholas Gao, Stephan Günnemann
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add Ewald message passing to the tested GNNs | SchNet, PaiNN, DimeNet++, and GemNet-T · OE62 | energy MAE: 16.1% | Improved |
| Add Ewald message passing to the tested GNNs | SchNet, PaiNN, DimeNet++, and GemNet-T · OC20 | energy MAE: 10.3% | Improved |
| Add Ewald message passing (ablation) | DimeNet++ · OE62 | energy MAE: 51.2 -> 46.9 | Improved |
| Vary Ewald frequency cutoff and voxel resolution to c_k = 0.4 Å⁻¹, Δ = 0.1 Å⁻¹ (ablation) | DimeNet++-Ewald · OE62 | energy MAE: 51.2 -> 47.2 | Improved |
| Vary Ewald frequency cutoff and voxel resolution to c_k = 1.6 Å⁻¹, Δ = 0.4 Å⁻¹ (ablation) | DimeNet++-Ewald · OE62 | energy MAE: 51.2 -> 47.3 | Improved |
| Add Ewald message passing (ablation) | SchNet · OE62 | energy MAE: 133.5 -> 83.0 | Improved |
| Vary Ewald frequency cutoff and voxel resolution to c_k = 0.2 Å⁻¹, Δ = 0.2 Å⁻¹ (ablation) | SchNet-Ewald · OE62 | energy MAE: 133.5 -> 85.9 | Improved |
| Vary Ewald frequency cutoff and voxel resolution to c_k = 0.8 Å⁻¹, Δ = 0.4 Å⁻¹ (ablation) | SchNet-Ewald · OE62 | energy MAE: 133.5 -> 82.8 | Improved |
| Add Ewald message passing (ablation) | SchNet · OC20 | energy MAE: 895 -> 836 | Improved |
| Change the periodic frequency indices to (1, 1, 2) (ablation) | SchNet-Ewald · OC20 | energy MAE: 895 -> 828 | Improved |
| Replace standard periodic filtering with radial filtering (ablation) | GemNet-T-Ewald · OC20 OOD | energy MAE: 416 -> 447 | Hurt |
| Replace standard periodic filtering with cuboid filtering (ablation) | GemNet-T-Ewald · OC20 OOD | energy MAE: 416 -> 436 | Hurt |
Closest papers by meaning
- Long-Short-Range Message-Passing: A Physics-Informed Framework to Capture Non-Local Interaction for Scalable Molecular Dynamics Simulation (ICLR 2024 · similarity 0.89)
- Neural Atoms: Propagating Long-range Interaction in Molecular Graphs through Efficient Communication Channel (ICLR 2024 · similarity 0.87)
- Learning Equivariant Non-Local Electron Density Functionals (ICLR 2025 · similarity 0.86)
- E2Former: An Efficient and Equivariant Transformer with Linear-Scaling Tensor Products (NeurIPS 2025 · similarity 0.86)
- The Importance of Being Scalable: Improving the Speed and Accuracy of Neural Network Interatomic Potentials Across Chemical Domains (NeurIPS 2024 · similarity 0.85)
- Higher-Rank Irreducible Cartesian Tensors for Equivariant Message Passing (NeurIPS 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
